BioPOS: Biologically Inspired Algorithms for POS Tagging, Proceedings of the 1st Workshop on Optimization Techniques for Language Technology

Abstract

In this paper we present a new biologically inspired approach to the part-of-speech tagging problem, based on particle swarm optimization. As far as we know this is the first attempt of solving this problem using swarm intelligence. We divided the part-of-speech problem into two subproblems. The first concerns the way of automatically extracting disambiguation rules from an annotated corpus. The second is related with how to apply these rules to perform the automatic tagging. We tackled both problems with particle swarm optimization. We tested our approach using two different corpora of English language and also a Portuguese corpus. The accuracy obtained on both languages is comparable to the best results previously published, including other evolutionary approaches.